Gender and diagnostic impact on everyday technology use: a differential item functioning (DIF) analysis of the Everyday Technology Use Questionnaire (ETUQ)

Abstract

Background: As the use of everyday technology is increasingly important for participation in daily activities, more in-depth knowledge of everyday technology use in relation to diagnosis and gender is needed. The purpose of this study was to investigate the stability of the perceived challenge of a variety of everyday technologies across different samples of varying diagnoses including both males and females.

Methods: This cross-sectional study used 643 data records from clinical and research samples, including persons with dementia or related disorders, acquired brain injury, intellectual disability, various mental or medical disorders, and adults without known diagnoses. The Everyday Technology Use Questionnaire, comprising 93 everyday technology artifacts and services (items) measuring the level of everyday technology challenge and relevance of and perceived ability to use these was used for data gathering. A two-faceted Rasch model in combination with differential item functioning (DIF) analyses were used for comparing item hierarchies across samples.

Results: Only three items (3.2%) demonstrated a clinically relevant DIF by gender, and nine items (9.7%) by diagnosis.

Discussion: The findings support a stable hierarchy of everyday technology challenge in home and community that can facilitate planning of an accessible and inclusive society from a technological departure point.

Implications for Rehabilitation

• The ability to manage everyday technology is increasingly important for participation in everyday activities at home and in the community for people with and without disabilities.

• This study demonstrates that differences in perceived challenges in using various everyday technologies across gender and diagnosis are minimal.

• The findings provide evidence of no or minor systematic bias in testing when using the Everyday Technology Use Questionnaire in clinical practice and research.

• Empirical knowledge about the perceived challenge of specific everyday technologies of people with variations in gender or diagnosis is still sparse, hence this study can inspire practice and future research.

Keywords:

• IADL/activities of daily living
• inclusive society
• aging
• cognitive impairment

Introduction

The use of technological artefacts and services such as mobile phones, internet banking and automatic ticket machines is increasing in homes as well as in public life and societal services. A huge variety of everyday technological (ET) artefacts and services, both digital and traditional non-digital technology, are currently in use in large populations internationally, and new technologies are continuously developed to further facilitate our lives, e.g., to promote inclusion and effective use of time, and also to compensate for disabilities [1,2]. Accordingly, the ability to manage everyday technology has become increasingly important for participation in everyday activities at home and in the community [1,3,4]. Even though technology is aimed to support everyday activities and participation in society it might for some people instead be a hindrance if the level of their ability and the technology’s level of challenge does not match [5]. The present study investigates this person–technology match through an analysis of the level of challenge across a variety of technological artefacts and services, based on the perceptions of people with a variety of diagnoses, when actually using these everyday technologies.

A concept closely related to “level of challenge” is “ease-of-use”, which also is an aspect of usability [6]. In order to be usable, a technological artefact or system must be managed easily in a number of ways; e.g., easy to learn and remember, and result in few errors [7]. Usability, in turn, is a prerequisite for the technology’s usefulness; i.e., how well it contributes to meeting the users’ goals in specific activities [6]. The latter aspect, focusing on the individually determined relevance of a piece of technology, is particularly important in the context of rehabilitation and health care, where the goal is to base interventions on the clients’ needs, yet usefulness has rarely been considered specifically in empirical research [8]. Similarly, although the concept ease-of-use is commonly accepted, less empirical knowledge exists on how easy or challenging people perceive those everyday technologies that are relevant in their lives to be. We know that if a technology is seen by the user to facilitate and support everyday life, he or she might be more motivated to use it. This might affect the user’s acceptance of the technology and also influence on its perceived level of challenge [9–11]. Thus, to be able to support peoples’ ability to use everyday technology we need information about their perceptions of everyday technology’s level of challenge.

In research targeting technology and people with disabilities, technology is most often seen as assets that provide means for inclusion. However, that is not always self-evident when it comes to people with multifaceted and cognitively disabling conditions such as dementia, mild cognitive impairment (MCI), acquired brain injury (ABI) or intellectual disability (ID). Our research has shown that technology also may be a hindrance in these people’s everyday lives and repeated studies have shown that people with more severe cognitive impairment have more difficulties to use technology than those with mild impairment, good recovery or no known cognitive impairment [12–15]. In a few studies, the challenge levels of common everyday technologies have also been investigated in samples with and without dementia or MCI [2,16–19] and samples with ABI [5]. Even though research has shown that people’s perceptions of a technology’s level of difficulty is related to the severity of cognitive decline, we do not yet know if people with different cognitively disabling diagnoses perceive everyday technologies to be similarly easy or challenging, or if it varies across diagnoses. Such knowledge would be important for example when planning for generic interventions on a societal level to support technology use in everyday life among people with and without disabilities in order to design a more accessible and inclusive society.

Moreover, gender issues are seldom addressed in research on everyday technology use. For example, in a recent review of the literature on gender and dementia care [20], technology use was not even mentioned. When gender is mentioned in relation to everyday technology use, findings have displayed more stereotypic notions (e.g., males being more interested in technology than females) [21]. A single-sided focus on gender differences can lead to strengthening of such stereotypes. More recent studies have however provided a more complex influence of gender onto everyday technology use [12,22]. Therefore, variations and similarities in the everyday technology level of challenge as perceived by women and men are equally important to identify.

In summary, more in-depth knowledge is needed about how diagnosis and gender relates to everyday technology use. Such knowledge can provide a basis for recommendations to further develop generic rehabilitation programs and contribute to the design of accessible and inclusive societies that could fit a variety of technology users. Hence, the purpose of this study was to investigate the stability of the person–everyday technology match across different clinical samples, by comparing how challenging a variety everyday technology artefacts and services were perceived to be in samples including both males and females with cognitive impairment due to a variety of diagnoses such as Alzheimer’s disease (AD), MCI ABI, ID and older adults with no known cognitive impairment. Our hypothesis was that not more than 5% of the everyday technologies investigated should demonstrate significant variations across gender or diagnosis, in order to provide generic profiles of everyday technology use.

Materials and methods

Participants and data gathering procedures

This cross-sectional study used 643 available data records from earlier recruited clinical and research samples, here divided into six different diagnostic groups; data records from persons with (I) dementia and related disorders (n = 156) (e.g., AD, and MCI), (II) ABI (n = 34) (e.g., stroke, traumatic brain injury), (III) ID (n = 114) (e.g., mild or moderate intellectual disability), (IV) mental health (n = 2), (V) medical disorders (n = 76) (e.g., orthopedic conditions), and (VI) older healthy adults without known cognitive impairments (n = 237). All data records were gathered in a Swedish context. Descriptive characteristics of the participants are shown in Table 1. All data records were gathered from participants that had earlier given informed consent to take part in this research, and ethical approvals were provided by the Ethical Committee at the Karolinska Institutet (2010/120–31/5). Demographic information of the included samples is presented in Table 1.

Table 1. Descriptive characteristics of the participants, number of relevant everyday technologies and ability to use everyday technologies generated from the Everyday Technology Use Questionnaire (ETUQ).

Group (n)	Sex, male/female (%)	Age, M (SD) min–max	Number of relevant everyday technologies (ETUQ), M (SD) min–max	Measure of the ability to use everyday technologies, expressed in logits (ETUQ), M (SD) min–max
Older healthy adults without known cognitive impairments (237)	93/144 (39/61)	60.2 (15.6) 13–92	62.4 (12.4) 20–91	59.5 (9.0) 46.3–89.7
Dementia and related disorders (156)	70/86 (45/55)	71.6 (10.3) 28–92	47.1 (13.7) 9–79	53.5 (5.6) 37.6–85.4
Acquired brain injury (34)	16/18 (47/53)	65.1 (17.5) 20–91	44.6 (13.5) 15–66	56.6 (5.6) 47.4–85.4
Intellectual disability (114)	62/52 (54/46)	38.8 (13.2) 18–67	47.5 (9.9) 19–68	49.3 (6.7) 30.0–61.2
Medical disorders (76)	24/52 (32/68)	75.9 (13.8) 27–99	47.6 (15.6) 17–81	69.7 (5.4) 44.9–69.7
Mental health^a (2)	1/1 (50/50)	49.0 (18.4) 36–62	56.5 (7.8) 51–62	58.1 (4.4) 55.0–61.2
No specified diagnosis^a (24)	6/18 (25/75)	77.5 (6.6) 64–90	34.8 (12.4) 9–58	56.0 (4.0) 49.0–65.4
Total (643)	272/371 (42/58)	61.8 (18.4) 13–99	52.3 (15.0) 9–91	55.3 (8.3) 30.0–89.7

^a Not included in analysis of DIF by diagnosis.

Instrument

In this study, we used the Everyday Technology Use Questionnaire (ETUQ) for data gathering. The ETUQ was developed based upon a Model of Human Occupation (MoHO) [23] and aims to identify the relevance of a large variety of everyday technologies, as well as the perceived ability to use these. The ETUQ comprises 93 technological artifacts and services, here called items (e.g., cell phones, coffee machines, Internet, automatic ticket machines), and is administered in a 30-to-45-min face-to-face interview. The items are grouped into eight different activity areas, including e.g., household occupations, accessibility, and transportation. For a more in-depth description of the ETUQ, see the study by Nygård et al. [14]. Initially, the ETUQ interview produces information about which everyday technologies are perceived as relevant or irrelevant in everyday life activities to the person. According to the ETUQ manual [24], an everyday technology is defined as relevant if it is available to the person and the person (i) uses it; (ii) has been using it or (iii) intends to start using it [42]. The perceived ability in the use of those everyday technologies that are relevant to the person is then registered by the interviewer. The reported difficulties are graded using a six-step scale (Table 2) [14]. The ETUQ generates two individual measures of everyday technology use: first, a count of the number of everyday technologies being relevant to the client. Secondly, a measure of the ability to use everyday technologies, generated by the use of Rasch analysis and provides a measure from 0–100, expressed in logits [13,14,25]. The ETUQ has in a number of studies with different samples demonstrated sound psychometric properties; more specifically acceptable internal scale validity, unidimensionality, person response validity, and discriminative validity in studies of adults with and without cognitive impairment [13,14,25].

Table 2. Description of the scale steps applied in the Everyday Technology Use Questionnaire (ETUQ) [24].

Categories	Description
A	Does not use the ET anymore or has not started to use it even if it is available and perceived as relevant
B	Always uses the ET together with another person
C	The ET is sometimes used together with another person
D	Uses the ET without another person, but with frequent/major perceived difficulties
E	Uses the ET without another person, but with minor perceived difficulties
F	Uses the ET without another person and without perceived difficulties

Data gathering procedures

All data were originally collected by experienced researchers, research assistants or clinical staff (all registered occupational therapists) who had been specifically trained by the developers of the ETUQ (Nygård and Rosenberg) in administering and scoring the ETUQ in a valid and reliable manner. Most ETUQ interviews took place in each participant’s home if the participant agreed on this, but interviews were also performed at workplaces, clinical sites or at local disability service centers.

Statistical analysis

A two-faceted Rasch rating scale model was chosen in this study to analyze the stability of the ETUQ item level of challenge across gender and diagnostic groups [26]. As the ETUQ was originally developed and evaluated using such mathematical models, it was a logical choice for this study. As Rasch models are also suitable for handling data where items may be missing at random, and it is expected that all ETUQ items are not relevant for all clients, we did not have to exclude any participant due to specific missing values by the use of this approach. This also allow us to compare people who may have different repertoires of relevant everyday technologies, as long as there are linkages between all people and all items in the data matrix. The WINSTEPS analysis software program, version 3.91.0 [27] was used to conduct the Rasch analyses. Such an analysis first converts raw item scores from ETUQ into equal-interval measures using a logarithmic transformation of the odds probabilities of the actual responses. The Rasch transformation simultaneously results in a measure of a person’s perceived ability to use everyday technologies, as well as an everyday technology item calibration measure reflecting the challenge of the specific everyday technology, along a similar continuum. The everyday technology calibration measures can then also be used to examine whether the items of a scale measure a one-dimensional construct, and if item challenges differs between subsamples [16,28].

Initially, a generic Rasch analysis was performed with data from all participants included (n = 643). This resulted in generic measures of ability to use everyday technologies for all participants and generic ETUQ item calibration measures for all items.

In order to evaluate the stability of the ETUQ item challenge hierarchy across gender and diagnostic groups we then split our sample into gender- and diagnosis-specific subsamples, and a number of differential item functioning (DIF) analyses were performed. We here decided to exclude the sample with cognitive impairment yet no specified diagnosis (n = 24) and the small sample with persons with mental disorders (n = 2) due to small sample sizes. In addition, given the impact of sample size and variability on standard errors of ETUQ item calibration measures across subsamples, differences between measures from two subsamples on a specific everyday technology could be artificially significant (due to relatively large or uneven sample sizes) but still not clinically relevant; therefore we decided to evaluate the size of the discrepancy between ETUQ item calibration measures using an earlier approach accounting for this potential bias, in which the standard error of the item calibration measure was set at 1.5 logit, indicating that an item difference must exceed 4.2 logit in order to clinically demonstrate a gender or diagnostic bias [29].

Finally, we also generated new ability measures for all participants based on the above generated gender- and diagnostic-specific ETUQ item hierarchies. We finally compared these specific measures generated from each of the gender- and diagnosis-specific ETUQ item hierarchy with the measures generated from the generic ETUQ item hierarchy for all participants using standardized z-comparisons (adjusting for the imprecision in each measure by the individual SEs); differential test functioning (DTF), A criterion was set that no more than 5% of our participants should differ significantly (z-values exceeding ±1.96) between the ability measures generated from the generic ETUQ scale versus the gender and diagnosis-specific ETUQ scale.

Results

When evaluating the ETUQ item calibrations for DIF by gender, three items of 93 (3.2%) demonstrated a clinically relevant DIF (#20 Iron relatively more challenging for men; #53 Electric grinding tool, and #54 Electric drill, relatively more challenging for women) (Table 2), which was less than our set criterion of 5%.

In the evaluation of DIF by diagnosis, nine items of 93 (9.7%) demonstrated a clinically relevant DIF, which was higher than our set criterion. Out of these nine items, five everyday technologies were demonstrated as relatively more challenging and four everyday technologies as relatively less challenging for persons with ID compared to other diagnostic groups. One item demonstrated DIF both in relation to gender and diagnosis (#20 Iron) (Table 3).

Table 3. ETUQ items demonstrating misfit in relation to gender and diagnosis (n = 643).

ETUQ item	DIF by gender^a	DIF by diagnosis^b	Direction of DIF^c
#10 Stove		X	ID > Med
#19 Trouser press		X	ID > Well ID > Dem
#20 Iron	X	X	Male > Female ID > Well
#33 Video player		X	Dem > ID Well > ID
#34 DVD player		X	Dem > ID
#36 Stereo		X	Dem > ID
#38 Tape player		X	ID > Well
#53 Electric grinding tool	X		Female > Male
#54 Electric drill	X		Female > Male
#56 Electric grass trimmer		X	Med > Dem
#81 Internet transactions		X	ID > Well

ID: intellectual disability; Med: medical conditions (e.g., orthopaedic conditions); Dem: dementia or related conditions (e.g., mild cognitive impairment); Well: older healthy adults without known cognitive impairment.

^a Differential item functioning by gender, as determined by an item calibration difference larger than 4.2 logits.

^b Differential item functioning by diagnosis, as determined by an item calibration difference larger than 4.2 logits.

^c Group A > Group B: The item is perceived as relatively more challenging for Group A as compared to Group B.

In the evaluation of potential DIF impact onto the person ability measures generated by the ETUQ items, none of the ability measures for the participants changed significantly; none of the ability measures changed more than 1.0 SE between the generic versus the gender-specific ETUQ scales (n = 643), and no ability measure out of 617 [excluding persons with no specified diagnosis (n = 24) and persons with Mental health issues (n = 2)] changed more than 2.0 individual SE.

So in summary, we concluded that the ETUQ demonstrated an acceptable number of item DIFs in relation to gender (n = 3), but higher than accepted in relation to diagnosis (n = 9) in this sample. The item DIFs detected were primarily caused by differences in the everyday technology hierarchy for people with ID in comparison to the other diagnostic groups. When evaluating the impact of these ETUQ item DIFs, they had no or minimal impact on the generated person measures. So even though some item DIFs are present in the ETUQ, valid comparisons can still be made across persons with variations in gender or diagnosis, using a generic ETUQ item scale. The generic ETUQ item hierarchy is presented in Figure 1. Participants perceiving having a higher ability to manage everyday technologies are placed higher on the continuum; participants perceiving less ability to manage everyday technologies are placed lower on the continuum. In a similar way, everyday technologies perceived as more challenging to manage are placed higher on the continuum; everyday technologies perceived as less challenging to manage are placed lower on the continuum.

Figure 1. Person-item map presenting the generic person ability measures of perceived ability to use ET from a heterogeneous sample (n = 643) and the measures of generic ETUQ item challenges based on selected representative items from the ETUQ (n = 31).

Discussion

The results of this study overall show that the levels of challenge of most everyday technologies were stable independent of gender and diagnosis as they only contributed to a clinically relevant difference on a limited number of everyday technologies. Moreover, these differences had no or minor impact on the generated person ability measures. The fact that the levels of challenge of the everyday technologies were overall similar between women and men supporting our hypothesis (only three out of 93 items demonstrated DIF) further highlights the importance of questioning stereotypic notions related to gender and everyday technology use. This was also found in earlier studies, where it was demonstrated both in people with ABI and among older people with and without MCI or AD that gender does not contribute to differences in ability to manage everyday technology [12,30]. The ETUQ evaluates the perceived difficulty only of items that are considered being relevant to the individual person and context. So if a technology is not accessible or available to the person, or if a technology is accessible but the person has never used it and has no intention to use it in the future, it is not scored in the ETUQ evaluation for that person. We argue that some of the stereotypic notions related to gender and technology are based upon technologies that may not always be considered relevant to a specific person, generation or context. If a technology is not considered relevant to a person and context it should not be impacting on generalized conclusions on everyday technology use (e.g., older males have problems using washing machines; older females have problems using smart phones). It is important that evaluation of technology use also takes into consideration if the technology is an integrated part of the person’s environment. If not, the gender conclusions drawn may be highly speculative and are not reflecting the real world experiences of people as technology users.

Also, the results indicate that people with different diagnoses related to cognition in most cases [84 items out of 93 (90.3%)] perceived everyday technologies to be relatively similar in challenge. The person-item map presented in Figure 1 can therefore be viewed as a relatively stable hierarchy that monitor the hierarchy of everyday technology from easier to harder that is generic for this large and diverse sample of people. This implies that such everyday technology hierarchy can be used for generic planning on a societal level to support the (ease of) use of everyday technology in home and community across populations with and without cognitive impairments from a variety of causes to improve cognitive accessibility and inclusion in society might be possible. The findings from this study provide empirical evidence which everyday technologies that are perceived as more difficult to manage (e.g., automatic check-in at airport; internet transactions) for a large and diverse sample of people. The more challenging everyday technologies as indicated in Figure 1 should therefore be the target for planning of interventions/modifications that would support a large and diverse sample of people. In line with the generic hierarchy provided in this study, another example of such planning could be universal design encouraging designers to consider the wide-ranging abilities of users to make everyday technologies easy to use for everyone to the greatest extent as possible, without the need of special design or adaptions [31].

It is also notable that even if the mean ability to use technology of a larger proportion of the population is placed higher than the mean of the everyday technology items’ levels of challenge (set by default at 50.0 logits), indicating that a majority of the sample do not perceive major problems in using most of the relevant everyday technologies (Table 3, Figure 1), still a considerable number of people perceive problems with many everyday technologies in home and community. This highlights the importance of user awareness in design to continue reducing the demand of everyday technology to make society more accessible and inclusive. By continuing to improve everyday technologies’ original designs, an equitable use of everyday technology for people with diverse abilities can be promoted. Moreover, in order to make everyday technology enable functioning and everyday activity performance, aspects of usability and usefulness are very important to highlight in such design processes [6,8,32]. The result also indicates that generic competent everyday technology use cannot be taken for granted in any of the examined diagnostic subsamples (not even among older people without known cognitive impairments). Everyday technology use may therefore need further individual attention in relation to enhancing an individual’s possibilities to participate in activities and situations requiring everyday technology use in home and community [33,34].

The detected ETUQ item DIFs by diagnosis seem to be explained by a somewhat different everyday technology item hierarchy found among people with ID, as eight out of nine ETUQ items demonstrating DIF by diagnosis were related to this sample. Among these eight items, five were perceived as more challenging for people with ID (Table 3). A possible reason for these items being identified as more challenging could reflect that people with ID not always are allowed or trusted to use these everyday technologies e.g., stove, iron, and internet transactions. In other words, these everyday technologies could have be given the lowest scores in ETUQ if they were available and the person wished to use them but has not started to use them. This would contribute to placing such everyday technologies at a higher level of challenge in the hierarchy as compared to the item hierarchies of other diagnostic groups. So the item DIF could here be caused by different trajectories reflecting the process of developing (or withdrawing from) everyday technology use among the various subsamples, where another group with a later in life onset of a progressive cognitive disorder (like dementia) may over time start perceive problems using a specific everyday technology, continue with support to use it, and finally stop using the specific everyday technology. The findings from this study may indicate that such trajectories in everyday technology integration/abandonment can be of importance to understand how people decide and approach incorporation or abandonment of technologies in their everyday life. Some initiatives have been taken to explore this phenomenon but research is still sparse [35,36].

Regarding the three media items that people with ID perceived relatively less challenging (Table 3) this might reflect more of a demographic difference; the group with ID has a lower mean age than the other groups and may therefore reflect more age-related issues, e.g., younger people (here those with ID) may have engaged more in activities that require more frequent use of such everyday technologies in comparison to older generations, which could influence the perceived level of challenge using such everyday technologies. It could also be so that older people demonstrate a higher level of visual [37] and/or auditory impairments [38], which also could influence the perceived difficulties with such everyday technologies. In earlier studies, age has not shown to be an aspect impacting on perceived or observed difficulties in everyday technology use [12,30], although younger people with ID perceived more everyday technologies being relevant to them in comparison to older people with ID [15]. However, most of these studies [12–14,19,30] have compared more homogeneous groups in relation to diagnosis and age span than this study. To be able to discover differences in everyday technology use due to age, further studies should consider wider age spans to also account for impact on different technology generations, as they may have different experiences, access and use of everyday technology [39].

Finally, it is important to note that when evaluating the impact of the ETUQ item DIFs, they demonstrated only minimal impact on the ETUQ person ability measures generated. So, even though some item DIFs are present in the ETUQ, valid comparisons can still be made across persons with variations in gender or diagnosis, using a generic ETUQ item scale. This contributes to validity evidence of internal construct and fairness in testing in relation to ETUQ, as people with variations in gender and/or diagnosis can be evaluated with the ETUQ with minimal bias on the ETUQ ability measures generated. The findings from this study therefore contributes to the growing evidence-base of the ETUQ as a valid tool to explore and measure everyday technology use among a variety of people [13,14,25,40,41].

There are some methodological issues that need to be reflected upon; relatively small sample sizes could impact on the precision of some of the everyday technology item calibration measures. Potential item DIFs can also occur within different diagnostic groups now combined into one subsample [e.g., people with left versus right cerebral vascular accident (CVA) were combined in subsample with ABI; people with frontal lobe dementia versus AD were combined in subsample dementia and related disorders], and also within the same group over time (e.g., people with MCI) [42–43]. The participants are also all gathered within a Swedish context. Although no or minimal ETUQ item DIF has been found when comparing matched samples across contexts [40,44], the generic hierarchy may still not be generalized to other contexts without caution. Finally, the use of self-report when gathering information about everyday technology use especially among people with cognitive impairments can be discussed as a limitation in this and other studies using the ETUQ [13,14,25,40,41].

Clinical implications

The results of this study contribute to an increased awareness and appreciation of similarities and differences in everyday technology use in different groups based on diagnoses or gender. The study findings support the hypothesis that everyday technologies can be ordered in a generic hierarchy from easier to harder that is applicable to a diverse sample, and overall astable across gender and diagnosis, with a few exceptions. This knowledge can support more generic user-centred approaches/initiatives on a societal level in promoting an inclusive design of everyday technologies for a diverse sample of people with and without disabilities, as similar perceived hierarchies in using various everyday technologies is empirically supported despite differences in gender or diagnosis. On an individual level, the large variations in ability to use everyday technologies presented in the findings indicate the importance of evaluating the ability to use everyday technology in rehabilitation processes (for individuals from various diagnostic groups with cognitive impairments) to identify those in need of support to overcome hindrances to participation in home and community activities requiring use of everyday technology.

Acknowledgements

The authors first of all want to thank the participants who generously demonstrated their management of everyday technology for us. The authors are also grateful to the registered OTs who collected and administered the data.

Disclosure statement

No potential conflict of interest was reported by the authors.

Additional information

Funding

The study was funded by the Swedish Research Council for Health Working Life and Welfare (FORTE), and the Swedish Research Council (VR).